If AI Can Watch My Groceries, Why Can’t It Help Moderate Nextdoor?
One of Nirav Tolia’s favorite topics on the speaking circuit is AI.
That got me thinking.
If AI is such a core part of Nextdoor’s strategy, why isn’t it being used more effectively to improve moderation?
AI isn’t new. I’ve watched it evolve over the past 25 years, and today it’s part of my everyday life.
Take a trip to my local Harris Teeter.
I walk through self-checkout, scan my groceries, bag my items, and leave. Cameras, sensors, and AI are constantly evaluating what’s happening. When something falls outside an expected pattern or decision matrix, the system alerts a human associate to step in.
It’s AI first.
Human review second.
That’s a scalable model.
Now compare that to what I continue to observe on Nextdoor.
The attached example contains snarky comments that remained visible two days after they were posted. This isn’t an isolated example; it’s part of a pattern I’ve documented during my ongoing Nextdoor experiment.
Why isn’t AI identifying conversations that are escalating into personal attacks or unconstructive exchanges and routing them to trained reviewers?
Instead, Nextdoor continues to rely heavily on a decentralized network of unpaid moderators. While many volunteer with good intentions, any moderation system benefits from consistent standards, quality assurance, and ongoing coaching.
To me, the current model feels like the inmates running the prison while the warden sits in the office, removed from the chaos.
If AI can help prevent mistakes at a grocery store checkout, surely it can help create a more consistent and constructive online community.
Join the discussion on NielFlamm.com.
The Great Uniter Is at It Again
It was only a matter of time.
Every Independence Day in the Lowcountry, fireworks become one of the hottest topics on Nextdoor. I’m all for spirited discussions—neighbors won’t always agree, and that’s healthy.
What I don’t understand is the inconsistency.
The thread I observed began about a week ago. Comments that, based on my own experience, I believe could have violated community standards have remained visible for five days.
That leaves me asking the same question I’ve been asking throughout my Nextdoor experiment:
Why are some comments allowed to remain while others result in moderation?
If moderation standards were applied consistently across neighborhoods, perhaps these situations would be less common.
My suggestion hasn’t changed.
Invest in AI to identify comments that may violate community standards using a clearly defined decision matrix. When the AI isn’t confident, route the content to trained human reviewers who receive ongoing coaching, calibration, and quality assurance.
That’s how many organizations deliver consistency.
Instead, Nextdoor continues to rely on a decentralized network of unpaid moderators. While many undoubtedly volunteer with good intentions, any moderation system benefits from oversight, feedback, and accountability.
Consistency builds trust.
Without it, users are left wondering whether the rules depend on the content—or on who’s reviewing it.
Join the discussion on NielFlamm.com.
My Nextdoor Experiment: When Moderation Leaves Questions Unanswered
As part of my ongoing Nextdoor experiment, I continue to observe discussions in my local community. Some of what I see is thoughtful. Some of it is disappointing.
With the 250th anniversary of the founding of the United States approaching, it’s encouraging to see neighbors celebrating a milestone in our country’s history. I’m grateful to have been born in the United States and appreciate the opportunities this country has provided.
What concerns me is when conversations move beyond the topic itself and begin singling out groups of people with broad stereotypes or negative generalizations.
Whether the target is based on race, ethnicity, religion, political affiliation, or another characteristic, allowing comments that stereotype an entire group can undermine the sense of community a platform aims to foster.
I included a screenshot of the discussion in this post so readers can judge the context for themselves.
That brings me back to a question I’ve asked repeatedly:
How are moderation decisions being made, and are they being applied consistently?
When users see some comments removed while others containing personal attacks or broad stereotypes remain, it’s natural to ask whether the moderation process is consistent.
I’ve suggested before that Nextdoor could strengthen its moderation model by combining AI-assisted detection with centralized quality assurance and regular moderator coaching. Regardless of the approach, consistency and transparency are important if users are expected to trust the process.
I’d like to see moderation that encourages constructive conversations while reducing comments that target or stereotype groups of people.
You can see the screenshots, read my full analysis, and join the discussion on NielFlamm.com/blog.
What has your experience been with online community moderation?
Day 16: Consistency Shouldn’t Depend on Who You Know
Every year around Independence Day, neighborhoods light up with discussions about fireworks and pets.
I understand why. It’s a topic that generates strong opinions.
What caught my attention wasn’t the topic itself—it was the comments that followed.
One commenter referred to another neighbor as “Karen.” By the way, I know an awesome Karen, Karen Romero and would recommend her for a leadership QA role in a heartbeat.
Another made a reference to Charles Darwin, implying that a species disappears because of poor choices—a comment clearly aimed at another person.
Yet those comments remained visible.
That leaves me asking the same question I’ve been asking throughout my Nextdoor experiment:
How are comments like these allowed to remain while other posts or comments are moderated or removed?
From my perspective, consistency is the issue.
Whether the inconsistency stems from differences in judgment, neighborhood dynamics, or something else, the result is the same: users may wonder whether the standards are being applied evenly.
My suggestion hasn’t changed.
Combine AI with centralized quality assurance and independent oversight. Let AI identify conversations that are becoming personal, and have trained reviewers apply the same standards across every neighborhood.
That seems like a better investment than expanding office space if the goal is to improve the user experience.
Meanwhile, today is Day 16 since I requested the Home Insurance Insights study referenced in a Nextdoor article.
I’ve now emailed Nirav Tolia, Jacob Chavis, and another member of the Customer Insights team.
Was it awkward?
A little.
But it also raises a broader question.
If this is the experience of someone making a straightforward request for information referenced in a public article, what should advertisers or users expect when they need help with a billing issue, campaign, or account problem?
Customer experience is built one interaction at a time.
I’d love to hear your thoughts.
Join the discussion on NielFlamm.com.
My Nextdoor Experiment: Six Days Later, the Conflict Is Still There
As part of my ongoing observation of the Nextdoor platform, I noticed something interesting this week.
By the way, as of writing this post, Nextdoor (NYSE: NXDR) is up approximately 4.5% to $2.32 per share. Markets move for many reasons, but I continue watching both the business and the user experience.
One of the biggest recurring topics in my neighborhood is e-bikes.
A post asking whether e-bikes should be permitted on sidewalks quickly turned into a debate. That wasn't surprising. Transportation, safety, and neighborhood etiquette are already topics that generate strong opinions.
What did surprise me was what happened next.
The discussion didn't cool off.
It escalated.
The following day, comments became increasingly personal, with one user being directly targeted rather than the discussion remaining focused on the issue itself.
And this wasn't a "my mistake" moment that was quickly corrected.
As of today, the post has remained active for six days.
This brings me back to the same question I've been asking for months.
If Nextdoor relies on unpaid community moderators, how is quality assurance ensured?
Who reviews moderator decisions?
Who coaches moderators?
Who ensures moderation is applied consistently across neighborhoods?
Every moderation system benefits from oversight. Customer service teams have quality monitoring. Call centers review interactions. Content moderation should be no different.
I've previously suggested a formal QA program with ongoing coaching, calibration, and AI-assisted detection to help identify escalating conversations before they become personal attacks.
Instead, I’m seeing what appears to be inconsistent moderation.
As a user and shareholder, I naturally question whether moderation standards are being applied consistently—or whether some users receive different treatment than others.
I'm genuinely interested in hearing from others.
Have you experienced inconsistent moderation or difficulty understanding how moderation decisions are made?
If so, I'd like to hear your story.
Feel free to connect with me here on LinkedIn, email me at niel@nielflamm.com, or call/text me at 843-212-6824.
One final observation: if you're a Nextdoor advertiser or user looking for phone-based customer support, you'll likely find no publicly available customer service number.
I'd love to hear your experiences.